{"url":"/sota/vcgbench-diverse-on-videoinstruct","task":{"name":"VCGBench-Diverse","url":"/task/vcgbench-diverse","note":null},"dataset":{"name":"VideoInstruct","url":"/dataset/videoinstruct"},"category":"Computer Vision","categories":["Computer Vision"],"category_note":null,"description":"Recognizing the limited diversity in existing video conversation benchmarks, we introduce VCGBench-Diverse to comprehensively evaluate the generalization ability of video LMMs. While VCG-Bench provides an extensive evaluation protocol, it is limited to videos from the ActivityNet200 dataset. Our benchmark comprises a total of 877 videos, 18 broad video categories and 4,354 QA pairs, ensuring a robust evaluation framework.\r\n\r\nThe evaluation is computed over five different aspects: \r\n\r\n1. Correctness of information \r\n\r\n2. Detail orientation \r\n\r\n3. Contextual understanding \r\n\r\n4. Temporal understanding \r\n\r\n5. Consistency. \r\n\r\nAdditionally, VCGBench-Diverse provides a breakdown of performance across three key aspects: \r\n\r\n1. Dense video captioning, which assesses the ability to generate detailed and accurate descriptions of the video content, \r\n\r\n2. Spatial understanding, which evaluates the capability to understand and describe the spatial relationships and settings within the video \r\n\r\n3. Reasoning, which tests the adeptness in inferring and explaining causal relationships and actions within the video.","description_from":"task","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","rank":"the archive's row order at snapshot; not re-ranked","rows_end_at":"2025-07-28","rows_withheld_as_spam":0,"metric_values":"the archive's strings, untouched"},"metrics":["mean","Correctness of Information","Detail Orientation","Contextual Understanding","Temporal Understanding","Consistency","Dense Captioning","Spatial Understanding","Reasoning"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"mean":null,"Correctness of Information":null,"Detail Orientation":null,"Contextual Understanding":null,"Temporal Understanding":null,"Consistency":null,"Dense Captioning":null,"Spatial Understanding":null,"Reasoning":null}},"counts":{"rows":6,"rows_with_code":6,"rows_with_paper_page":6,"rows_dated":6,"rows_using_additional_data":0},"rows":[{"rank_in_archive_order":1,"model":"VideoGPT+","metrics":{"Consistency":"2.59","Contextual Understanding":"2.81","Correctness of Information":"2.46","Dense Captioning":"1.38","Detail Orientation":"2.73","Reasoning":"3.63","Spatial Understanding":"2.80","Temporal Understanding":"1.78","mean":"2.47"},"uses_additional_data":false,"paper_date":"2024-06-13","paper":"/paper/videogpt-integrating-image-and-video-encoders","paper_url":"https://arxiv.org/abs/2406.09418v1","paper_title":"VideoGPT+: Integrating Image and Video Encoders for Enhanced Video Understanding","code":"https://github.com/mbzuai-oryx/videogpt-plus","n_code_links":1,"syntology":{"n_ran":6,"n_unverified":2,"n_samples":8,"n_pointer_only_licence":8}},{"rank_in_archive_order":2,"model":"Chat-UniVi","metrics":{"Consistency":"2.36","Contextual Understanding":"2.66","Correctness of Information":"2.29","Dense Captioning":"1.33","Detail Orientation":"2.56","Reasoning":"3.59","Spatial Understanding":"2.36","Temporal Understanding":"1.56","mean":"2.29"},"uses_additional_data":false,"paper_date":"2023-11-14","paper":"/paper/chat-univi-unified-visual-representation","paper_url":"https://arxiv.org/abs/2311.08046v3","paper_title":"Chat-UniVi: Unified Visual Representation Empowers Large Language Models with Image and Video Understanding","code":"https://github.com/pku-yuangroup/chat-univi","n_code_links":4,"syntology":null},{"rank_in_archive_order":3,"model":"VideoChat2","metrics":{"Consistency":"2.27","Contextual Understanding":"2.51","Correctness of Information":"2.13","Dense Captioning":"1.26","Detail Orientation":"2.42","Reasoning":"3.13","Spatial Understanding":"2.43","Temporal Understanding":"1.66","mean":"2.20"},"uses_additional_data":false,"paper_date":"2023-11-28","paper":"/paper/mvbench-a-comprehensive-multi-modal-video","paper_url":"https://arxiv.org/abs/2311.17005v4","paper_title":"MVBench: A Comprehensive Multi-modal Video Understanding Benchmark","code":"https://github.com/opengvlab/ask-anything","n_code_links":3,"syntology":{"n_ran":7,"n_unverified":3,"n_samples":10,"n_pointer_only_licence":0}},{"rank_in_archive_order":4,"model":"BT-Adapter","metrics":{"Consistency":"2.27","Contextual Understanding":"2.59","Correctness of Information":"2.20","Dense Captioning":"1.03","Detail Orientation":"2.62","Reasoning":"3.62","Spatial Understanding":"2.35","Temporal Understanding":"1.29","mean":"2.19"},"uses_additional_data":false,"paper_date":"2023-09-27","paper":"/paper/one-for-all-video-conversation-is-feasible","paper_url":"https://arxiv.org/abs/2309.15785v2","paper_title":"BT-Adapter: Video Conversation is Feasible Without Video Instruction Tuning","code":"https://github.com/farewellthree/BT-Adapter","n_code_links":1,"syntology":{"n_ran":1,"n_unverified":3,"n_samples":4,"n_pointer_only_licence":0}},{"rank_in_archive_order":5,"model":"VTimeLLM","metrics":{"Consistency":"2.35","Contextual Understanding":"2.48","Correctness of Information":"2.16","Dense Captioning":"1.13","Detail Orientation":"2.41","Reasoning":"3.45","Spatial Understanding":"2.29","Temporal Understanding":"1.46","mean":"2.17"},"uses_additional_data":false,"paper_date":"2023-11-30","paper":"/paper/vtimellm-empower-llm-to-grasp-video-moments","paper_url":"https://arxiv.org/abs/2311.18445v1","paper_title":"VTimeLLM: Empower LLM to Grasp Video Moments","code":"https://github.com/huangb23/vtimellm","n_code_links":1,"syntology":{"n_ran":5,"n_unverified":6,"n_samples":11,"n_pointer_only_licence":11}},{"rank_in_archive_order":6,"model":"Video-ChatGPT","metrics":{"Consistency":"2.06","Contextual Understanding":"2.46","Correctness of Information":"2.07","Dense Captioning":"0.89","Detail Orientation":"2.42","Reasoning":"3.60","Spatial Understanding":"2.25","Temporal Understanding":"1.39","mean":"2.08"},"uses_additional_data":false,"paper_date":"2023-06-08","paper":"/paper/video-chatgpt-towards-detailed-video","paper_url":"https://arxiv.org/abs/2306.05424v2","paper_title":"Video-ChatGPT: Towards Detailed Video Understanding via Large Vision and Language Models","code":"https://github.com/mbzuai-oryx/video-chatgpt","n_code_links":2,"syntology":null}],"since_archive":{"claim":"Results that newer papers report for their own method, placed here by Syntology. A model pointed at the cell in the paper's own table; the number was read from that cell and checked against this leaderboard's metric, dataset, split and scale; an independent check that saw this leaderboard's other rows and every other leaderboard on the same dataset accepted it. Not reviewed by the paper's authors or by the archive's editors, and not ranked against the archive rows.","extraction_file_present":true,"measurement":{"test_papers":883,"papers_with_output":881,"judged_true":108,"judged":110,"wilson95_lower":0.9361,"measured_on":"2026-09-24","frozen_commit":"0e3de0df94"},"measurement_note":"blind adjudication of accepted entries on a held-out split of archive papers, rules frozen before the test","coverage":{"sentence":"Syntology has checked 6,885 of the 9,623 papers on this site that are newer than the archive; results from the others appear after they are checked.","complete":false,"papers_newer_than_archive":9623,"papers_checked":6885,"papers_extracted_not_yet_verified":0,"boards_without_verdict":2,"papers_not_yet_extracted":2737},"order":"newest first by month (arXiv date, else the arXiv-id month), then arXiv id descending","columns":[],"entries":[]},"syntology":{"read_at":"2026-09-25T09:33:49+00:00","claim":"Per row: N of M harvested code samples from that row's paper executed on a synthesized fixture; the other M-N are unverified. Not a reproduction of the row's number; not a correctness claim. n_pointer_only_licence counts samples the site points at rather than redistributes (a licence axis, independent of ran/unverified).","rows_with_graph_line":4,"rows_with_any_sample_ran":4,"distinct_papers_with_graph_line":4,"distinct_papers_with_any_sample_ran":4,"samples_over_distinct_papers":{"n_ran":19,"n_unverified":14,"n_samples":33,"n_pointer_only_licence":19,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":19,"n_unverified":14,"n_samples":33,"n_pointer_only_licence":19,"note":"row-weighted: a paper behind several rows is counted once per row; inflated relative to samples_over_distinct_papers by design, kept for readers summing the per-row syntology blocks"}}}